Listening Deepfake Detection on ListenForge 1.0 (test)
97.24AUCMANet
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MANetModality=Audio-Visual, Retraining Status=Proposed method (trained on target dataset)2026.04 | 97.24 | 89.74 | |
| AVTFDModality=Audio-Visual, Retraining Status=with retraining2026.04 | 93.19 | 89.28 | |
| MRDFModality=Audio-Visual, Retraining Status=with retraining2026.04 | 90.32 | 90.58 | |
| CViTModality=Visual, Retraining Status=with retraining2026.04 | 90.16 | 87.96 | |
| MesoNetModality=Visual, Retraining Status=with retraining2026.04 | 89.31 | 88.8 | |
| XceptionModality=Visual, Retraining Status=with retraining2026.04 | 88.24 | 87.23 | |
| XceptionModality=Visual, Retraining Status=without retraining2026.04 | 62.02 | 57.23 | |
| CViTModality=Visual, Retraining Status=without retraining2026.04 | 56.64 | 52.15 | |
| AVADModality=Audio-Visual, Retraining Status=without retraining2026.04 | 55.18 | 41.78 | |
| AVTFDModality=Audio-Visual, Retraining Status=without retraining2026.04 | 54.3 | 44.19 | |
| MRDFModality=Audio-Visual, Retraining Status=without retraining2026.04 | 45.98 | 50.47 | |
| MesoNetModality=Visual, Retraining Status=without retraining2026.04 | 43.28 | 46.99 |